original article activation of insulin-reactive cd8 t...

Activation of Insulin-Reactive CD8 T-Cells forDevelopment of Autoimmune DiabetesF. Susan Wong,

1Lai Khai Siew,

1Gwen Scott,

1Ian J. Thomas,

1Stephen Chapman,

1Christophe Viret,

2

and Li Wen3

OBJECTIVE—We have previously reported a highly diabeto-genic CD8 T-cell clone, G9C8, in the nonobese diabetic (NOD)mouse, specific to low-avidity insulin peptide B15-23, and cellsresponsive to this antigen are among the earliest islet infiltrates.We aimed to study the selection, activation, and development ofthe diabetogenic capacity of these insulin-reactive T-cells.

RESEARCH DESIGN AND METHODS—We generated a T-cellreceptor (TCR) transgenic mouse expressing the cloned TCRV�18/V�6 receptor of the G9C8 insulin-reactive CD8 T-cell clone.The mice were crossed to TCRC��/� mice so that the majority ofthe T-cells expressed the clonotypic TCR, and the phenotype andfunction of the cells was investigated.

RESULTS—There was good selection of CD8 T-cells with apredominance of CD8 single-positive thymocytes, in spite ofthymic insulin expression. Peripheral lymph node T-cells had anaïve phenotype (CD44lo, CD62Lhi) and proliferated to insulinB15-23 peptide and to insulin. These cells produced interferon-�and tumor necrosis factor-� in response to insulin peptide andwere cytotoxic to insulin peptide–coated targets. In vivo, theTCR transgenic mice developed insulitis but not spontaneousdiabetes. However, the mice developed diabetes on immuniza-tion, and the activated transgenic T-cells were able to transferdiabetes to immunodeficient NOD.scid mice.

CONCLUSIONS—Autoimmune CD8 T-cells responding to alow-affinity insulin B-chain peptide escape from thymic negativeselection and require activation in vivo to cause diabetes.Diabetes 58:1156–1164, 2009

Type 1 diabetes is a complex, multifactorial dis-ease in which genetic factors interact with envi-ronmental modifiers to give rise to immuneabnormalities leading to pancreatic �-cell dam-

age and destruction. Both CD4 and CD8 T-cells have amajor role in pathogenesis of type 1 diabetes. There are anumber of autoantigens that are recognized by CD8 T-cells, including proinsulin (PI) (1), islet-specific glucose-6-phosphatase catalytic subunit–related protein (IGRP) (2),dystrophia myotonica kinase (3), and glutamic acid decar-boxylase (4) in the nonobese diabetic (NOD) animal model

of diabetes. In HLA-A2 transgenic mice, both PI and IGRPhave also been shown to be important targets (5,6).Furthermore, emerging evidence suggests that CD8 T-cellsin humans also recognize these antigens (7–9), althoughtheir role in pathogenesis is currently unknown.

In humans, CD8 T-cells are present in islet infiltrates atthe time diabetes develops (10). CD8 T-cells were predom-inant in the islet infiltrate when diabetes recurred within 6weeks after transplantation of a hemipancreas from nor-mal, identical nondiabetic twins into their diabetic cotwins(11). In NOD mouse studies, it is well established that CD8T-cells play an important role both in early events leadingto insulitis and diabetes (12,13) as well as in the finaleffector stage (14) of diabetes development.

PI is an important autoantigen in diabetes, and it hasbeen suggested that it is the “prime” autoantigen (15)recognized by CD8 T-cells, CD4 T-cells, and autoantibod-ies. There is clear evidence that PI is expressed in thethymus (16) in humans and in mice (17), which influencesthe expression of autoreactive T-cells to PI/insulin. Inhumans, although the major histocompatibility complex(MHC) is the most important genetic susceptibility factor,the second most important genetic region (IDDM2) is theinsulin 5�VNTR region. This controls expression of PI inthe thymus and the pancreas (18). Mice have two types ofinsulin, produced as proinsulin 1 (PI1), mainly in thepancreas, and proinsulin 2 (PI2) in the thymus and pan-creas. Studies using individual proinsulin PI1�/� andPI2

�/�

knockout mice, when backcrossed to the NODbackground, showed that PI1�/� mice have a reducedincidence of diabetes (19), whereas in PI2�/� mice, diabe-tes is accelerated with 100% developing diabetes (19,20).In addition, insulin autoantibody production is increasedin PI2�/� mice, and spleen cells from young PI2�/� ani-mals have an increased ability to transfer diabetes (20).Furthermore, when PI2 was overexpressed in the thymus(and on peripheral antigen-presenting cells [APCs]) on theMHC class II promoter (PI2tg), the incidence of diabeteswas decreased (21,22). This suggests that PI2 is importantfor both central and peripheral tolerance for islet �-cells.

We have previously isolated a highly pathogenic CD8T-cell clone (G9C8) from the islets of young, pre-diabeticNOD mice that is capable of very rapidly causing diabetes(5–10 days) upon adoptive transfer to young or irradiatednondiabetic NOD mice (23). The autoantigen is an insulinB-chain peptide (amino acids 15–23 [B15-23]) (1). Theepitope is common to both mouse insulins (and is alsoconserved in human insulin) and is restricted by MHC-Kd.The insulin peptide is recognized by a small population ofcells present in the very early stages of the islet infiltrate,identified using a Kd–B15-23 tetramer (1) and is found inmice aged �5 weeks (1,24). Subsequently, other specific-ities such as IGRP become increasingly dominant (24–26),and this insulin-specific population becomes a smaller

From the 1Department of Cellular and Molecular Medicine, University ofBristol, Bristol, U.K.; the 2Centre de Physiopathologie de Toulouse Purpan,INSERM, U563-BAT A CHU Purpan, Toulouse, France; and the 3Section ofEndocrinology, Yale University School of Medicine, New Haven,Connecticut.

Corresponding author: Susan Wong, [email protected] 17 June 2008 and accepted 4 February 2009.Published ahead of print at http://diabetes.diabetesjournals.org on 10 Febru-

ary 2009. DOI: 10.2337/db08-0800.© 2009 by the American Diabetes Association. Readers may use this article as

long as the work is properly cited, the use is educational and not for profit,and the work is not altered. See http://creativecommons.org/licenses/by-nc-nd/3.0/ for details.

The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked “advertisement” in accordance

with 18 U.S.C. Section 1734 solely to indicate this fact.

ORIGINAL ARTICLE

1156 DIABETES, VOL. 58, MAY 2009

percentage of the infiltrate as the disease progresses (1).To further study the selection and activation of this earlypopulation of CD8 T-cells, we have generated a T-cellreceptor (TCR) transgenic mouse in which the T-cellsexpress the receptor of the highly pathogenic G9C8 clone.Unlike other CD8 TCR transgenic mice that recognizehigher-affinity MHC binding peptides of IGRP (2,27), G9C8T-cells have low avidity, and G9 transgenic mice do notdevelop spontaneous diabetes. However, transgenic G9cells can be highly diabetogenic after activation both invitro and in vivo.

RESEARCH DESIGN AND METHODS

Generation of insulin-reactive TCR transgenic mice. Insulin-specific TCRtransgenic mice were generated by isolating TCR genomic DNA from G9C8cloned T-cells (23), which have previously been shown to have specificreactivity to amino acids 15–23 of the insulin B-chain (1). TCR� (V�18S1 J�18C�)- and � (V�6S1 D�1.1 J�2.3 C�1)-chain DNA was purified and cloned intopT�cass and pT�cass constructs (28), respectively, kindly provided by D.Mathis. The cloned constructs were injected directly into NOD ova to generateindependent TCR� and � founder lines in the Yale Diabetes EndocrinologyResearch Center (DERC) transgenic core (R.A. Flavell, Yale School ofMedicine), which were intercrossed to produce �� TCR transgenic mice(G9NOD). Three independent founder lines each of TCR� (lines 17, 22, and 24)and � (lines 1, 2, and 3) were generated, and � lines 17, 22, and 24 werecrossed with each of � lines 2 and 3. Results with each of these lines wassimilar in terms of selection and functional phenotype (data not shown), and,thus, TCR� line 22 and TCR� line 2 were chosen for all subsequent studies.

To generate repertoire-restricted �� TCR transgenic mice, G9C��/�.NODmice were generated by crossing the �� TCR transgenic mice to NOD.C��/�

mice. The NOD.C��/� mice were generated by backcrossing TCR.C��/�

mice (B6/129 background; kindly provided by A.C. Hayday) (29) �20generations to NOD/Caj mice in the laboratory of R.S. Sherwin (YaleUniversity). G9RAG�/�.NOD mice were generated by crossing the ��TCR transgenic mice to NOD.RAG2�/� mice, generated by backcrossingRAG2�/� mice (kindly provided by D. Schatz, Yale University) to NOD/Cajmice �12 generations. To generate G9C��/�.RIP-B7.1 NOD mice, theRIP-B7.1 mice were 10 generations backcrossed to NOD.C��/� mice andintercrossed to generate NOD.C��/�RIP-B7.1 mice and then crossed withthe G9C��/�.NOD mice to generate G9C��/�.RIP-B7.1 NOD mice. NOD.scidmice were originally obtained from The Jackson Laboratories but werebred for many years at Yale University. The mice were housed inmicroisolators or scantainers in barrier rooms. All procedures wereperformed in accordance with the protocols approved by the U.K. HomeOffice or institutional animal care and use committee at Yale University.Flow cytometry. Single-cell suspensions of 0.1–1 � 106 cells were stained forvarious cell surface markers (rat anti-murine CD8; CD4, CD62L, and CD44).FcRs were blocked with anti-CD16 (2.4G2; provided by the late C.A. Janeway),and cells were stained for surface markers. Intracellular staining was per-formed on cells cultured for 15 h with differing concentrations of peptide, forinterferon (IFN)-� and tumor necrosis factor (TNF)-�, after cell surfacestaining with anti-CD8 following the BD Bioscience protocol. They wereanalyzed by flow cytometry using a FACSCalibur, and results were analyzedusing CellQuest (BD Bioscience) or FlowJo (Tree Star) software. All antibod-ies were purchased from BD Bioscience, except anti-murine CD8-allophyco-cyanin (Caltag Laboratories).

Cells were stained with Kd-B15-23-PE tetramer or control Kd-listeriolysin(LLO) 91-99-PE tetramer (obtained from the National Institutes of HealthTetramer Facility) for 1 h at 4°C. Anti–CD8–fluorescein isothiocyanate wasadded for the last half hour. The cells were then washed and analyzed by flowcytometry.

Enriched CD8 T-cells from splenocytes of G9C��/�.NOD mice werelabeled with 5 mol/l carboxyfluorescein succinimidyl ester (CFSE) for 10min at 37°C and washed with ice-cold RPMI-1640 with 5% FCS and then with0.9% normal saline. Ten million cells were resuspended in 0.9% normal salineand transferred intravenously to 4- or 6-week-old NOD mice. The spleen,pancreatic, axillary, and mesenteric lymph nodes were removed 6–10 daysposttransfer, and CFSE dilution was analyzed by flow cytometry after stainingwith anti–CD8-allophycocyanin and anti–CD69-phycoerythrin.Proliferation and cytotoxicity assays: proliferation. Insulin-reactive CD8T-cells from 5- to 8-week-old G9C��/�.NOD mice were cocultured with bonemarrow–derived dendritic cells and B15-23 peptide or influenza hemagglutininpeptide IYSTVASSL (control) for peptide responses in RPMI complete me-dium. Similarly, soluble insulin or keyhole limpet hemocyanin (control) was

used as antigen for protein cross-presentation responses. CD8 T-cells werepurified from splenocytes of G9C��/�.NOD mice using positive selectionbeads (Miltenyi Biotec) with �95% purity. After 3 days of incubation, cellswere pulsed with [3H]thymidine for 14 h to determine proliferation.Cytotoxicity. CD8 T-cells purified from 5- to 8-week-old G9C��/�.NOD mice,used directly ex vivo or prestimulated for 24 h with 1 g/ml insulin B15-23peptide, were harvested and further cocultured with 104 51Cr-sodium chro-mate (Amersham)-labeled P815 cells together with B15-23 insulin peptide atan effector-to-target (E:T) ratio of 10:1 for 16 h. Specific lysis was calculatedas [(cytotoxic release � min)/(max � min)] � 100%, where the minimalrelease (min) corresponded to the spontaneous lysis, and the maximal lysiscorresponded to lysis induced by addition of Triton-X100 (max).In vivo immunization. G9C��/�.NOD mice were immunized intraperitone-ally with 50 g B15-23 peptide together with different concentrations of CpG1826 oligonucleotides (Coley pharmaceuticals). Nine to twenty-one days later,a further injection of the same dose was given. Controls were injected witheither 50 g B15-23 peptide or the different doses of CpG alone.Diagnosis of diabetes. Diabetes was detected by initially testing for glyco-suria and confirmed by blood glucose measurements �13.9 mmol/l/l.Adoptive transfer of diabetes. A total of 5–7 � 106 purified CD8 T-cellsfrom G9C��/�.NOD mice, either taken directly ex vivo or preactivated byCpG-matured bone marrow–derived dendritic cells and insulin B15-23 pep-tide, were washed, resuspended in a volume of 200 l of normal saline, andtransferred intravenously to 6-week-old NOD.scid mice, 3-week-old NODmice, and 6-week-old NOD mice. CD4 T-cells purified either from young6-week-old mice (107) or from mice that had become diabetic (4–10 � 106)were transferred intravenously to 3- to 4-week-old G9C��/�.NOD mice.Histology. Pancreata were harvested and fixed, and immunohistochemistrywas performed as previously described (30). Insulitis was measured in 6-, 9-,12-, 15-, 20-, and 25-week-old mice (n 3–4 each). Sections were stained withhematoxylin and eosin (Vector Laboratories) and scored for insulitis.

RESULTS

TCR transgenic cells are selected in the thymus. Thetransgenic TCR in G9NOD mice is expressed in boththymus and the periphery, and there is an increasedproportion of CD8 T-cells and reduced CD4 T-cells com-pared with nontransgenic NOD mice (Fig. 1A–E). Theallelic exclusion at the TCR� chain locus is poor (Fig. 1Cand F), and other TCR� chains can be selected as shownby a high frequency of expression of the TCR� chains 2, 3,8, and 11, which were identified by flow cytometry. Similarresults were obtained from different TCR�� transgeniclines (combination of mice from three different TCR�founder lines with two different TCR� founder lines; datanot shown). Thus, to study the transgenic TCR�� of theoriginal G9C8 T-cell clone, we selected one TCR�� trans-genic line and crossed the G9NOD mice with TCRC��/�

and RAG2�/� mice in order to study the characteristics ofthe repertoire-restricted TCR�� transgenic cells.

Despite thymic expression of preproinsulin 2 (31), theG9C��/� and G9RAG�/� transgenic cells are clearly se-lected in the thymus (Fig. 2A and F). There is peripheralexpansion of the T-cells in the spleen and lymph nodes(illustrated in Fig. 2B and C and Fig. 2G and H, respec-tively), and, as expected, the cells express TCRV�6(shown in Fig. 2D and I), and the majority of cellsexpressing CD8 are also positive for Kd-insB15-23 tetramerstaining (Fig. 2E).G9C��/� TCR transgenic cells respond to antigen byproliferation, cytokine production, and cytotoxicity.The T-cells demonstrate antigen-specific cytotoxicity di-rectly ex vivo (shown in Fig. 3A), but their ability to kill invitro targets is enhanced by nearly two logs after culturingthe cells with antigen, indicating that the cells are rela-tively naïve when derived directly ex vivo from the trans-genic mouse (Fig. 3A). They produce IFN-� and TNF-� inresponse to antigenic peptide (Fig. 3B). As expected, theyproliferate to insulin peptide (Fig. 3C) and also respond toinsulin protein cross-presented by bone marrow–derived

F.S. WONG AND ASSOCIATES

DIABETES, VOL. 58, MAY 2009 1157

dendritic cells (shown in Fig. 3D), indicating that thepeptide is naturally processed and presented. Thus, thetransgenic G9 cells maintain the immunological functionof their parental G9C8 clone.TCR transgenic cells cause insulitis but not diabetes.In view of the highly pathogenic nature of the originalG9C8 T-cell clone, we tested whether the mice developedinsulitis. We found that in the G9C��/� TCR transgenicmice, a few islets have a large infiltration by CD8 T-cellsbut also by B-cells and CD4 T-cells (online appendixFig. 1 [available at http://diabetes.diabetesjournals.org/cgi/content/full/db08-0800/DC1]). However, infiltration isonly seen after 15 weeks of age, and even then most of theislets do not become infiltrated (islets examined from miceat 6, 9, 12, 15, 20, and 25 weeks of age) (insulitis scoresshown in online appendix Table 1). Neither the G9C��/�

TCR mice (30 mice observed until 35 weeks of age) nor theG9RAG�/� TCR mice developed spontaneous diabetes,similar to the three lines of G9NOD mice observed thatwere not on the C��/� background.Role of CD4 T-cells. As can be seen from Fig. 2, althoughthere are fewer CD4 T-cells in the G9C��/�.NOD micecompared with the G9NOD mice (shown in Fig. 1), thequestion arises as to whether these CD4 T-cells contain a

regulatory T-cell population that prevented the G9C��/

�.NOD mice from developing diabetes. To examinewhether the CD4 T-cell population inhibited the CD8T-cells in vitro, proliferation assays were performed thatshowed that G9C��/�.NOD CD4 T-cells can proliferate tothe insulin B15-23 peptide but at a lower level. Further-more, they do not inhibit the CD8 T-cells from proliferat-ing when they are added to CD8 T-cells in the proliferationassay but rather enhance the proliferation (shown in Fig.4A). However, to test further whether these cells werehaving an inhibitory effect in vivo, we depleted both totalCD4 T-cells (n 6) using anti-CD4 antibody andCD4�CD25� T-cells (n 3) using anti-CD25 antibodyinjected four times over a period of 6 weeks starting at 4weeks of age. The experiments were controlled by injec-tion of rat isotype control antibody (n 3). Efficacy ofdepletion was shown by testing peripheral blood justbefore the next injection, and at each time point in miceinjected with anti-CD4 or anti-CD25, the CD4 orCD4�CD25� T-cells were �1% in the peripheral blood(data not shown). None of the mice in any of the groups,observed until 20 weeks of age, developed diabetes. Thus,the lack of diabetes did not appear to be related to CD4regulatory T-cells.

FIG. 1. Comparison of thymocytes and splenocytes from G9NOD transgenic mice and nontransgenic NOD mice. A–C: Cells from G9NOD transgenicmice. D–F: Cells from nontransgenic NOD mice. A and D: Thymocytes single-positive CD4 and anti-CD8. B and E: Splenocytes stained withanti-CD4 and anti-CD8. C and F: Splenocytes stained with anti-CD8 and a pool of anti-V�2, -3, -8, and -11. The results are representative of 15transgenic mice in more than five independent experiments. The mean (SD) values for the cell populations are as follows: A: CD4 3.46 (1.71); CD85.89 (2.28). B: CD4 17.38 (4.91); CD8 20.35 (3.77). C: CD8V�2, -3, -8, -11 negative 20.0 (4.1); CD8V�2, -3, -8, -11 positive 3.8 (1.9). D: CD4 10.2(1.9); CD8 4.7 (0.7). E: CD4 40.63 (5.6); CD8 17.07(1.9). F: CD8V�2, -3, -8, -11 negative 15.36 (1.03); CD8V�2, -3, -8, -11 positive 1.85 (0.08).

CD8 T-CELLS AND AUTOIMMUNE DIABETES


To further examine whether CD4 T-cell help was re-quired, G9C��/�.NOD mice were infused with either puri-fied CD4 T-cells from young NOD mice or with CD4 T-cellsfrom diabetic NOD mice. None of the G9C��/�.NOD micegiven CD4 cells from young NOD mice developed diabetes,whereas diabetes occurred in G9C��/�.NOD mice trans-ferred with CD4 T-cells from the diabetic NOD spleens(shown in Fig. 4B). Thus, activated but not nonactivatedpolyclonal CD4 T-cells are able to promote diabetes in theG9C��/�.NOD mice.TCR transgenic cells become highly diabetogenic af-ter activation in vitro or in vivo. The G9 cells in TCRtransgenic mice express a naïve phenotype and function,and the G9C��/�.NOD mice do not develop spontaneousdiabetes. Therefore, we sought to test whether activationof the insulin-reactive transgenic T-cells, other than byadding activated CD4 T-cells, could cause diabetes. Weinjected the mice with insulin B15-23 peptide together withtype B CpG oligonucleotides and showed that this ap-proach promoted very rapid diabetes development (Fig.5A). This was a specific response requiring peptide, asCpG alone did not induce disease (data not shown).Similarly, the CpG was required to activate the APCs, asinjection with peptide alone also did not induce disease(data not shown). The combined CpG and peptide injec-tion activated G9 transgenic T-cells, and their activationprofile shows downregulation of CD62L and upregulationof CD44 (illustrated in Fig. 5B). On development of diabe-tes, most islets were seen to have extensive infiltration anddestruction, whereas little or no insulitis was seen in theislets of mice that were immunized with either CpG orpeptide alone (shown in Fig. 5C).

Furthermore, the G9C��/�.NOD CD8 T-cells could be-come potent diabetogenic T-cells after activation withbone marrow–derived dendritic cells that are matured byCpG. Following a 3-day culture with CpG-matured bone

marrow dendritic cells in the presence of insulin B15-23peptide, the G9C��/�.NOD CD8 transgenic cells wereable to transfer diabetes rapidly without the aid of CD4T-cells into 6-week-old NOD.scid mice (Fig. 6), recapit-ulating the phenotype of the original CD8 T-cell clones(23). G9C��/�.NOD CD8 T-cells that were not preacti-vated did not transfer diabetes when a similar number ofnonactivated G9C��/�.NOD CD8 T-cells were transferredand observed for 12 weeks after transfer (data not shown).This suggests that activation increases the functional avidityof G9 cells (shown by the data in Fig. 3A), and this enhancestheir cytotoxic function and diabetogenicity.Activation of insulin-reactive CD8 T-cells by increas-ing antigen presentation within islets. We have previ-ously suggested that expression of the costimulator B7-1on the islets is able to activate low-avidity T-cells andcause accelerated diabetes (30). To test whether thecoexpression of RIP-B7.1 could induce diabetes inG9C��/�.NOD mice, we crossed the G9C��/�.NOD miceto RIP-B7.1.C��/� mice that had been generated for thispurpose. In the presence of B7.1 in the pancreas, theG9C��/�.NOD mice developed spontaneous diabetes(shown in Fig. 7). To determine that this was a periph-eral effect and not related to changes in selection of theG9TCR T-cells, we had previously demonstrated that thenaïve G9TCR CD8 T-cells from G9C��/�.NOD micecould readily transfer diabetes to NOD.scid mice thatexpressed RIP-B7.1 (30), indicating that the naïveG9TCR CD8 T-cells can be activated in situ in the isletsand become highly pathogenic. These results supportthe data presented earlier that increasing the functionalavidity of G9TCR T-cells through activation can pro-mote diabetogenicity of transgenic G9 T-cells.Activation of insulin-reactive CD8 T-cells in vivo inNOD mice. To investigate the homing and activation ofinsulin-reactive G9 CD8 T-cells in vivo, we used CFSE-

β

FIG. 2. Phenotype of G9C��/�.NOD mice and G9RAG�/�.NOD mice. G9C��/�.NOD thymocytes (A), splenocytes (B), and PLN cells (C) werestained with anti-CD4 and -CD8. D: PLN cells were stained with anti-V�6 and CD8. E: Gated CD8 T-cells were stained with Kd-B15-23 (bold line)and control Kd-LLO 91–99 (dotted line) tetramer, respectively. G9RAG�/�.NOD thymocytes (F), splenocytes (G), and PLN cells (H) were stainedwith anti-CD4 and -CD8. I: PLN cells were stained with anti-V�6 and CD8. The results represent three independent experiments. The mean (SD)values for the distribution of cell populations are as follows: A: CD4 1.43 (0.42); CD8 5.1 (1.6). B: CD4 8.28 (1.35); CD8 21.14 (4.62). C: CD4 6.28(0.57); CD8 52.18 (9.26). F: CD4 0.85 (0.44); CD8 2.93 (1.22). G: CD4 2.21 (2.16); CD8 33.42 (6.94). H: CD4 1.18 (0.66); CD8 68.7 (13.32).



labeled, purified CD8 T-cells from G9C��/�.NOD micetransferred into 4- to 6-week-old NOD mice and analyzedCFSE dilution in different lymphoid tissues. We showedthat the G9 CD8 T-cells proliferated, as shown by thedecrease in CFSE staining, mainly in the pancreatic lymphnodes (PLNs) but not in the spleen or other lymph nodes.More importantly, the dividing G9 CD8 T-cells expressearly activation marker CD69 (Fig. 8). This indicates thatNOD mice as young as 4 weeks of age have the capabilityof activating G9-like cells in the pancreatic lymph nodes.However, this single transfer of nonactivated cells into 4-to 6-week-old NOD mice did not result in acceleration ofdiabetes, when the mice were followed to 15 weeks (theage by which acceleration of disease would be observed).

DISCUSSION

Our study has shown that G9 TCR transgenic CD8 T-cellsthat react to an insulin B-chain peptide B15-23, one of theearliest antigenic epitopes to which CD8 T-cells respond(at 3–4 weeks of age), are selected in the thymus. Oncereleased in the periphery, they can be stimulated tobecome highly pathogenic. The G9 cells are activated bythe B15-23 peptide, which has low-affinity binding to theMHC-Kd, and, interestingly, as expected from a naturallyprocessed and presented peptide, transgenic G9 cells canbe stimulated by whole insulin protein cross-presented bydendritic cells. Although G9C��/� mice do not spontane-ously develop diabetes, they do so when naturally acti-vated polyclonal CD4 T-cells from diabetic NOD mice are

transferred, indicating a requirement for CD4 T-cells.Moreover, when transgenic G9 cells are fully activated invitro, they gain potent diabetogenicity and cause diabetesin immunodeficient NOD.scid mice, without additionalCD4 T-cell help, recapitulating the phenotype of the pa-rental clone cells. Importantly, these cells can also beactivated in vivo and cause rapid diabetes development byproviding exogenous insulin peptide in the presence ofimmunostimulatory CpG. A similar effect can be achievedif the costimulatory molecule B7.1 is expressed within theislets.

Why are these low-avidity T-cells present in the NODmouse and where are the cells stimulated in vivo? Re-cently, a multitude of peripheral antigens were shown tobe expressed in medullary epithelial cells of the thymusdue to the autoimmune regulator gene Aire (32). High-avidity T-cells are removed at the cortico-medullary junc-tion, protecting against autoimmunity, and deficiency ofAire leads to widespread multiorgan autoimmune disease(32). However, some autoreactive T-cells still escape fromthe thymus, and, using model antigens, it has been shownthat low-avidity T-cells are not deleted in the thymus(33,34). PI2 is expressed in the thymus (17,35), and evenwith the thymic expression, G9-like low-avidity insulin-reactive T-cells can clearly escape from negative selection.Furthermore, in our studies, low-avidity insulin-specificCD8 T-cells are found in the periphery and are not subjectto peripheral deletion. As with other islet autoantigen-reactive reactive CD8 (36) and CD4 (37) T-cells in the NOD

µµ

µµ

γα

FIG. 3. Functional responses of purified CD8 T-cells from G9C��/�.NOD mice. A: Cytotoxic assay of insulin-reactive CD8 T-cells directly ex vivo(E) from G9C��/�.NOD mice compared with precultured cells (F) in response to increasing B15-23 peptide concentrations at an E:T ratio of 10:1.B: Intracellular IFN-� and TNF-� production in response to increasing concentrations of insulin B15-23 peptide after a 40-h culture. C:3H-Thymidine incorporation proliferation assay of splenocytes from G9C��/�.NOD mice in response to B15-23 peptide and to influenzahemagglutinin (flu HA) control peptide. F, Ins B15-23; E, flu HA. D: 3H-Thymidine incorporation proliferation assay of splenocytes fromG9C��/�.NOD mice in response to cross-presented insulin protein or control KLH. The results were from one of four independent experiments.



mouse, G9 cells are activated and proliferate preferentiallyin the pancreatic lymph nodes compared with the spleenand other lymph nodes, including the mesenteric lymphnodes. However, our earlier study and those of others(1,3,24) demonstrated that the numbers of these cellsremain small in lymphoid tissue, but they are measurablein the islets from an early age. It is possible that only in theislets is sufficient antigen presented to these cells for theirexpansion, as a relatively high concentration of the low-affinity peptide is required to stimulate them. These cellscan be sufficiently activated in the pancreatic draininglymph nodes, possibly by dendritic cells that stimulatethem to traffic to the islets where a higher concentration ofthe insulin peptide within the islets induces G9 cell expan-sion and damage to the islets in the early phases. However,other specificities that may be of higher avidity, such asIGRP-reactive CD8 T-cells, expand more and overtakethese cells in number and percentage representation in theislet infiltrate in later phases of the disease development.

Why do these TCR transgenic mice not develop diabetesspontaneously? In the NOD mouse, different islet antigen–specific T-cells are likely to be involved in early phases ofthe autoimmune process, and it is likely that each mayamplify the autoimmune response of the others by produc-ing interleukin-2 locally, which may also increase cells byhomeostatic expansion (38). Insulin, as a �-cell–specificautoantigen, is very different from other �-cell–associatedantigens. In addition to MHC class I–restricted antigenpresentation by the �-cells, insulin is secreted by islet�-cells in response to high glucose from the portal system

and circulates in the body at low concentrations, depen-dent on ambient glucose. The peptide recognized by theinsulin-specific T-cells in this study has a low affinity ofbinding to H2-Kd molecules (39). The widely studied 8.3TCR transgenic T-cells recognize an epitope of IGRP206-214, an enzyme within �-cells with high avidity (40).Much less is known of the other autoantigen, dystrophiamyotonica kinase, which stimulates the CD8 T-cells fromthe TCR transgenic mouse AI-4 (3). Both of these micedevelop diabetes spontaneously, although 8.3TCR trans-genic mice develop more diabetes in the presence of CD4T-cells (27), whereas this is not a requirement for the AI-4transgenic mouse (41). In contrast, for CD4 TCR trans-

µ

FIG. 4. The role of CD4 T-cells. A: 3H-Thymidine incorporation prolif-eration assay of CD8, CD4, and mixed CD8 and CD4 splenocytes fromG9C��/�.NOD mice in response to B15-23 peptide G9C��/�. B: Adop-tive transfer of purified CD4 T-cells of splenocytes from diabetic NODintravenously into 4-week-old G9C��/�.NOD mice (n � 16). No diabe-tes occurred in mice adoptively transferred with CD4 T-cells purifiedfrom young NOD mice (n � 5). P < 0.018 when both groups werecompared. The mice were tested weekly for glycosuria and diabetesconfirmed by blood glucose of >13.9mmol/l.

FIG. 5. Activated CD8 T-cells from G9C��/�.NOD mice cause diabetes. A:G9C��/�.NOD mice were immunized with B15-23 peptide (50 �g) togetherwith CpG oligonucleotides at different doses (two injections, 9–21 daysapart). Diabetes was seen after the second injection in the majority ofmice and the graph illustrates the mice, following the second injection,immunized with peptide plus 65 �g CpG (F), peptide plus 100 �g CpG (E),and peptide plus 195 �g CpG (‚). Two of four mice immunized withpeptide, and high-dose CpG (195 �g) actually developed diabetes beforethe second injection and the remaining two mice are shown in the graph.None of the control mice injected with peptide alone (n � 11) or thevarious doses of CpG alone (n � 13) developed diabetes (data notshown). P < 0.0001 when all groups were compared. B: Flow cytometry ofspleen cells from a nonimmunized G9C��/�.NOD mouse, gated on CD8T-cells, shown in the thin line, compared with cells from a diabetic mouseafter CpG and peptide immunization, shown in the thick line. The upper

panel shows upregulation of CD44 and the lower panel illustrates down-regulation of CD62L. C: Histology showing sections from diabetic G9C��/

�.NOD mouse immunized with CpG and peptide (top row), CpG alone(middle row), and peptide alone (bottom row), stained with anti-CD8,anti-CD4, and anti-B220 (B-cells). (A high-quality representation of thisfigure is available in the online issue.)



genic mice, which also have varying incidence of diabetes,the highly pathogenic BDC2.5 cells develop spontaneousdiabetes when expressed on a C57BL/6 genetic back-ground, but the incidence of diabetes is much lower on theNOD genetic background (42), although it is sharplyincreased when regulatory T-cells are removed by expres-sion on a NOD.scid genetic background. Similarly, the 4.1TCR transgenic mice recognizing an unknown islet antigendevelop diabetes a little earlier on a RAG2�/� geneticbackground (27), and the 12-4.1 TCR transgenic mice,recognizing an insulin peptide only, develop diabetes on aRAG2�/� genetic background (43). Not surprisingly, thesemice have varying phenotypes, differing between CD4 andCD8 transgenic T-cells that are likely to reflect importantaspects of the biology of the native cells within the normalmice and also interactions with other cells of the immunesystem, emphasizing the need to study more than onespecificity of pathogenic T-cell.

The fact that the G9TCR transgenic mice do notdevelop spontaneous diabetes is highly reminiscent ofthe LCMVgp/p14 double transgenic mouse, in which atransgenic TCR that recognizes a peptide of lymphocyticchoriomeningitis virus (LCMV) glycoprotein was coex-pressed with the target LCMVgp as an autoantigen in thepancreas using the rat insulin promoter (RIP) (44). In thatmodel, the TCR transgenic cells were “ignorant” anddiabetes did not occur spontaneously but could be in-duced to occur on infection with LCMV or administrationof the LCMVgp peptide with lipopolysaccharide (44). Here,

we show that this mechanism can operate for a naturallyoccurring autoantigen (insulin) and low-avidity (G9) T-cells. In the NOD mouse, which develops diabetes spon-taneously, infection is clearly not the strong stimulusnecessary for activation of low-avidity cells, and, indeed,many infections prevent diabetes in these mice. However,there may be endogenous stimuli of the innate immunesystem that could activate low-avidity cells (45,46), leadingto further immune activation in both diabetes and otherautoimmune diseases. Indeed, we have shown that activa-tion, using a ligand for TLR9, can stimulate the insulin-reactive cells to become highly potent cytotoxic T-cells.The challenge will be to identify how this process occursin vivo.

What relevance does this low-avidity T-cell have indiabetes? A recent study (15) showed that diabetes wasinhibited if NOD mice were deficient in both PI1 and PI2but expressed a transgenic insulin that was mutated atposition 16 of the B-chain. The alteration of B16, in fact,will undoubtedly affect both pathogenic CD4 cells recog-nizing insulin B9-23 (47) as well as pathogenic CD8 T-cellsrecognizing insulin B15-23 (23). This B16 mutation over-laps the CD8 epitope B15-23 and is the major MHC-Kd

binding residue for the B15-23 epitope (39). Its effects inprotecting against diabetes were likely to be linked toremoving not only the CD4 epitope affecting diabetogenicCD4 cells but also the CD8 epitope, abolishing the abilityof G9-like CD8 T-cells to recognize and bind naturalB15-23, a weak binder to the MHC-Kd (39). Thus, althoughthe study by Nakayama et al. (15) did not distinguisheffects on CD4 insulin– versus CD8 insulin–reactive T-cells, CD8 T-cells are highly likely to be affected (48,49).G9-like cells are present in the islets at a very early stageof islet infiltration (1), and others have shown that reac-

FIG. 6. Adoptive transfer of purified activated G9C��/�.NOD CD8T-cells into NOD.scid mice. G9C��/�.NOD CD8 T-cells were purifiedand cultured with matured bone marrow–derived dendritic cells to-gether with 0.5 �g/ml insulin B15-23 peptide for 3 days. After washing,the cells were intravenously injected into 6-week-old NOD.scid mice(n � 8). Mice were tested daily for glycosuria and diabetes confirmedby blood glucose >13.9mmol/l.

FIG. 7. Diabetes incidence in G9C��/�.RIP-B7.1 NOD mice. G9C��/�.NODmice that also expressed RIP-B7.1 transgene (F, n � 18) were comparedwith nontransgenic G9C��/�.NOD mice (E, n � 20) and observed fordiabetes. Mice were tested weekly for glycosuria and diabetes confirmedby blood glucose >13.9 mmol/l (P < 0.0001).

FIG. 8. Insulin-reactive CD8 T-cells proliferate only in PLNs but not inthe spleen, MLN, or ALN when adoptively transferred into NOD mice.CFSE-labeled purified G9C��/�.NOD CD8 T-cells were injected iv into4-week-old NOD mice and the lymphoid tissue extracted and cellsanalyzed in PLNs (A) and spleen (B). The cells were gated on CD8T-cells, and CFSE-labeled cells, costained with anti-CD69, are shownfrom PLNs (A) and spleen (B). Percentage proliferation of divided cellsis shown after gating on CD8 T-cells from PLNs (C) and spleen (D).Proliferation in MLN and ALN was similar to spleen (data not shown).All groups were compared by ANOVA P � 0.014.



tivity to B15-23 is the earliest CD8 antigenic specificity sofar identified when compared with the more dominantIGRP-reactive T-cells (24). Cytotoxic T-cells specific forIGRP were not detected in PItg NOD mice that weretolerant to PI, suggesting that immune responses againstPI are required for IGRP-specific T-cells to develop (50).The current study reports insulin-reactive CD8 transgenicT-cells can cause rapid diabetes after activation, and thisstudy does not address whether insulin is a “prime autoan-tigen” for CD8 T-cells in NOD mice. Further work isongoing to establish whether these cells require “help”from CD4 insulin–reactive cells or whether other specific-ities can provide this.

In conclusion, we have shown that a CD8 T-cell respon-sive to a natural peptide of insulin, a major autoantigen inautoimmune diabetes, is selected in the thymus and canbecome highly pathogenic once activated in the periphery,causing diabetes. These T-cells are “ignorant” in a naïvestate, but this can be readily overcome by stimuli thatinclude direct activation or activation of antigen-present-ing cells, either in pancreatic lymph nodes or in the islets.CD4 T-cells are likely to play a significant role, and this iscurrently under investigation. Further experiments aim toidentify the endogenous stimulators that facilitate the“switch” of naïve and “ignorant” G9 cells into potent�-cell–destructive effector cells in the NOD mouse. Ourstudy provides an important model not only for the studyof immunopathogenic role of insulin-specific CD8 T-cellsin autoimmune diabetes but also for better design ofinsulin-based immunotherapy.

ACKNOWLEDGMENTS

These studies were funded by the Wellcome Trust, theJuvenile Diabetes Research Foundation, and the NationalInstitute of Diabetes and Digestive and Kidney Diseases(DERC, transgenic core, and NOD mouse genetic core).Tetramers were obtained from the National Institutes ofHealth Core Facility.

No potential conflicts of interest relevant to this articlewere reported.

In memory of Charles A. Janeway, Jr., without whosesupport in the early stages, this work would not havebeen possible. We thank Dr. Andrew Herman and theUniversity of Bristol Cellular and Molecular MedicineFlow Cytometry Facility for assistance in these studiesand Dr. Alexander Chervonsky for helpful comments onthe manuscript. We also thank Ewan Basterfield, JennyRadcliffe, Louise Phelon, and Craig Husher for expertcare of the animals.

REFERENCES

1. Wong FS, Karttunen J, Dumont C, Wen L, Visintin I, Pilip IM, Shastri N,Pamer EG, Janeway CA Jr. Identification of an MHC class I-restrictedautoantigen in type 1 diabetes by screening an organ-specific cDNA library[see comments]. Nat Med 1999;5:1026–1031

2. Lieberman SM, Evans AM, Han B, Takaki T, Vinnitskaya Y, Caldwell JA,Serreze DV, Shabanowitz J, Hunt DF, Nathenson SG, Santamaria P,DiLorenzo TP. Identification of the beta cell antigen targeted by a prevalentpopulation of pathogenic CD8� T cells in autoimmune diabetes. Proc NatlAcad Sci U S A 2003;100:8384–8388

3. Lieberman SM, Takaki T, Han B, Santamaria P, Serreze DV, DiLorenzo TP.Individual nonobese diabetic mice exhibit unique patterns of CD8� T cellreactivity to three islet antigens, including the newly identified widelyexpressed dystrophia myotonica kinase. J Immunol 2004;173:6727–6734

4. Quinn A, McInerney MF, Sercarz EE. MHC class I-restricted determinantson the glutamic acid decarboxylase 65 molecule induce spontaneous CTLactivity. J Immunol 2001;167:1748–1757

5. Jarchum I, Baker JC, Yamada T, Takaki T, Marron MP, Serreze DV,DiLorenzo TP. In vivo cytotoxicity of insulin-specific CD8� T-cells inHLA-A*0201 transgenic NOD mice. Diabetes 2007;56:2551–2560

6. Takaki T, Marron MP, Mathews CE, Guttmann ST, Bottino R, Trucco M,DiLorenzo TP, Serreze DV. HLA-A*0201-restricted T cells from humanizedNOD mice recognize autoantigens of potential clinical relevance to type 1diabetes. J Immunol 2006;176:3257–3265

7. Mallone R, Martinuzzi E, Blancou P, Novelli G, Afonso G, Dolz M, Bruno G,Chaillous L, Chatenoud L, Bach JM, van Endert P. CD8� T-cell responsesidentify beta-cell autoimmunity in human type 1 diabetes. Diabetes 2007;56:613–621

8. Baker C, Petrich de Marquesini LG, Bishop AJ, Hedges AJ, Dayan CM,Wong FS. Human CD8 responses to a complete epitope set from prepro-insulin: implications for approaches to epitope discovery. J Clin Immunol2008;28:350–360

9. Skowera A, Ellis RJ, Varela-Calvino R, Arif S, Huang GC, Van-Krinks C,Zaremba A, Rackham C, Allen JS, Tree TI, Zhao M, Dayan CM, Sewell AK,Unger W, Drijfhout JW, Ossendorp F, Roep BO, Peakman M. CTLs aretargeted to kill beta cells in patients with type 1 diabetes throughrecognition of a glucose-regulated preproinsulin epitope. J Clin Invest2008;118:3390–3402

10. Itoh N, Hanafusa T, Miyazaki A, Miyagawa J, Yamagata K, Yamamoto K,Waguri M, Imagawa A, Tamura S, Inada M, et al. Mononuclear cellinfiltration and its relation to the expression of major histocompatibilitycomplex antigens and adhesion molecules in pancreas biopsy specimensfrom newly diagnosed insulin-dependent diabetes mellitus patients. J ClinInvest 1993;92:2313–2322

11. Sibley RK, Sutherland DE, Goetz F, Michael AF. Recurrent diabetesmellitus in the pancreas iso- and allograft: a light and electron microscopicand immunohistochemical analysis of four cases. Lab Invest 1985;53:132–144

12. Wicker LS, Leiter EH, Todd JA, Renjilian RJ, Peterson E, Fischer PA,Podolin PL, Zijlstra M, Jaenisch R, Peterson LB. �2-Microglobulin–defi-cient NOD mice do not develop insulitis or diabetes. Diabetes 1994;43:500–504

13. Serreze DV, Gallichan WS, Snider DP, Croitoru K, Rosenthal KL, Leiter EH,Christianson GJ, Dudley ME, Roopenian DC. MHC class I-mediated antigenpresentation and induction of CD8� cytotoxic T-cell responses in autoim-mune diabetes–prone NOD mice. Diabetes 1996;45:902–908

14. Kay TW, Parker JL, Stephens LA, Thomas HE, Allison J. RIP-beta 2-micro-globulin transgene expression restores insulitis, but not diabetes, in beta2-microglobulin null nonobese diabetic mice. J Immunol 1996;157:3688–3693

15. Nakayama M, Abiru N, Moriyama H, Babaya N, Liu E, Miao D, Yu L,Wegmann DR, Hutton JC, Elliott JF, Eisenbarth GS. Prime role for aninsulin epitope in the development of type 1 diabetes in NOD mice. Nature2005;435:220–223

16. Pugliese A. Central and peripheral autoantigen presentation in immunetolerance. Immunology 2004;111:138–146

17. Derbinski J, Gabler J, Brors B, Tierling S, Jonnakuty S, Hergenhahn M,Peltonen L, Walter J, Kyewski B. Promiscuous gene expression in thymicepithelial cells is regulated at multiple levels. J Exp Med 2005;202:33–45

18. Vafiadis P, Bennett ST, Todd JA, Nadeau J, Grabs R, Goodyer CG,Wickramasinghe S, Colle E, Polychronakos C. Insulin expression in humanthymus is modulated by INS VNTR alleles at the IDDM2 locus. Nat Genet1997;15:289–292

19. Moriyama H, Abiru N, Paronen J, Sikora K, Liu E, Miao D, Devendra D,Beilke J, Gianani R, Gill RG, Eisenbarth GS. Evidence for a primary isletautoantigen (preproinsulin 1) for insulitis and diabetes in the nonobesediabetic mouse. Proc Natl Acad Sci U S A 2003;100:10376–10381

20. Thebault-Baumont K, Dubois-Laforgue D, Krief P, Briand JP, Halbout P,Vallon-Geoffroy K, Morin J, Laloux V, Lehuen A, Carel JC, Jami J, Muller S,Boitard C. Acceleration of type 1 diabetes mellitus in proinsulin 2-deficientNOD mice. J Clin Invest 2003;111:851–857

21. French MB, Allison J, Cram DS, Thomas HE, Dempsey-Collier M, Silva A,Georgiou HM, Kay TW, Harrison LC, Lew AM. Transgenic expression ofmouse proinsulin II prevents diabetes in nonobese diabetic mice. Diabetes1997;46:34–39

22. Jaeckel E, Lipes MA, von Boehmer H. Recessive tolerance to preproinsulin2 reduces but does not abolish type 1 diabetes. Nat Immunol 2004;5:1028–1035

23. Wong FS, Visintin I, Wen L, Flavell RA, Janeway CA Jr. CD8 T cell clonesfrom young nonobese diabetic (NOD) islets can transfer rapid onset ofdiabetes in NOD mice in the absence of CD4 cells. J Exp Med 1996;183:67–76

24. Trudeau JD, Kelly-Smith C, Verchere CB, Elliott JF, Dutz JP, Finegood DT,Santamaria P, Tan R. Prediction of spontaneous autoimmune diabetes in



NOD mice by quantification of autoreactive T cells in peripheral blood.J Clin Invest 2003;111:217–223

25. Amrani A, Verdaguer J, Serra P, Tafuro S, Tan R, Santamaria P. Progres-sion of autoimmune diabetes driven by avidity maturation of a T- cellpopulation. Nature 2000;406:739–742

26. Anderson B, Park BJ, Verdaguer J, Amrani A, Santamaria P. PrevalentCD8(�) T cell response against one peptide/MHC complex in autoimmunediabetes. Proc Natl Acad Sci U S A 1999;96:9311–9316

27. Verdaguer J, Schmidt D, Amrani A, Anderson B, Averill N, Santamaria P.Spontaneous autoimmune diabetes in monoclonal T cell nonobese diabeticmice. J Exp Med 1997;186:1663–1676

28. Kouskoff V, Signorelli K, Benoist C, Mathis D. Cassette vectors directingexpression of T cell receptor genes in transgenic mice. J Immunol Methods1995;180:273–280

29. Philpott KL, Viney JL, Kay G, Rastan S, Gardiner EM, Chae S, Hayday AC,Owen MJ. Lymphoid development in mice congenitally lacking T cellreceptor alpha beta-expressing cells. Science 1992;256:1448–1452

30. Thomas IJ, Petrich de Marquesini LG, Ravanan R, Smith RM, Guerder S,Flavell RA, Wraith DC, Wen L, Wong FS. CD86 has sustained costimulatoryeffects on CD8 T cells. J Immunol 2007;179:5936–5946

31. Derbinski J, Schulte A, Kyewski B, Klein L. Promiscuous gene expressionin medullary thymic epithelial cells mirrors the peripheral self. NatImmunol 2001;2:1032–1039

32. Anderson MS, Venanzi ES, Klein L, Chen Z, Berzins SP, Turley SJ, vonBoehmer H, Bronson R, Dierich A, Benoist C, Mathis D. Projection of animmunological self shadow within the thymus by the aire protein. Science2002;298:1395–1401

33. von Herrath MG, Dockter J, Oldstone MB. How virus induces a rapid orslow onset insulin-dependent diabetes mellitus in a transgenic model.Immunity 1994;1:231–242

34. Zehn D, Bevan MJ. T cells with low avidity for a tissue-restricted antigenroutinely evade central and peripheral tolerance and cause autoimmunity.Immunity 2006;25:261–270

35. Throsby M, Homo-Delarche F, Chevenne D, Goya R, Dardenne M, PleauJM. Pancreatic hormone expression in the murine thymus: localization indendritic cells and macrophages. Endocrinology 1998;139:2399–2406

36. Zhang Y, O’Brien B, Trudeau J, Tan R, Santamaria P, Dutz JP. In situ betacell death promotes priming of diabetogenic CD8 T lymphocytes. J Immu-nol 2002;168:1466–1472

37. Turley S, Poirot L, Hattori M, Benoist C, Mathis D. Physiological beta celldeath triggers priming of self-reactive T cells by dendritic cells in a type-1diabetes model. J Exp Med 2003;198:1527–1537

38. Kamimura D, Bevan MJ. Naive CD8� T cells differentiate into protectivememory-like cells after IL-2 anti IL-2 complex treatment in vivo. J Exp Med2007;204:1803–1812

39. Wong FS, Moustakas AK, Wen L, Papadopoulos GK, Janeway CA Jr.Analysis of structure and function relationships of an autoantigenic

peptide of insulin bound to H-2K(d) that stimulates CD8 T cells ininsulin-dependent diabetes mellitus. Proc Natl Acad Sci U S A 2002;99:5551–55562

40. Han B, Serra P, Yamanouchi J, Amrani A, Elliott JF, Dickie P, Dilorenzo TP,Santamaria P. Developmental control of CD8 T cell-avidity maturation inautoimmune diabetes. J Clin Invest 2005;115:1879–1887

41. Graser RT, DiLorenzo TP, Wang F, Christianson GJ, Chapman HD,Roopenian DC, Nathenson SG, Serreze DV. Identification of a CD8 T cellthat can independently mediate autoimmune diabetes development in thecomplete absence of CD4 T cell helper functions. J Immunol 2000;164:3913–3918

42. Zucchelli S, Holler P, Yamagata T, Roy M, Benoist C, Mathis D. Defectivecentral tolerance induction in NOD mice: genomics and genetics. Immu-nity 2005;22:385–396

43. Jasinski JM, Yu L, Nakayama M, Li MM, Lipes MA, Eisenbarth GS, Liu E.Transgenic insulin (B:9-23) T-cell receptor mice develop autoimmunediabetes dependent upon RAG genotype, H-2g7 Homozygosity, and insulin2 gene knockout. Diabetes 2006;55:1978–1984

44. Garza KM, Chan VS, Ohashi PS. T cell tolerance and autoimmunity. RevImmunogenet 2000;2:2–17

45. Kim HS, Han MS, Chung KW, Kim S, Kim E, Kim MJ, Jang E, Lee HA, YounJ, Akira S, Lee MS. Toll-like receptor 2 senses beta-cell death andcontributes to the initiation of autoimmune diabetes. Immunity 2007;27:321–333

46. Tian J, Avalos AM, Mao SY, Chen B, Senthil K, Wu H, Parroche P, DrabicS, Golenbock D, Sirois C, Hua J, An LL, Audoly L, La Rosa G, Bierhaus A,Naworth P, Marshak-Rothstein A, Crow MK, Fitzgerald KA, Latz E, KienerPA, Coyle AJ. Toll-like receptor 9-dependent activation by DNA-containingimmune complexes is mediated by HMGB1 and RAGE. Nat Immunol2007;8:487–496

47. Wegmann DR, Norbury-Glaser M, Daniel D. Insulin-specific T cells are apredominant component of islet infiltrates in pre-diabetic NOD mice. EurJ Immunol 1994;24:1853–1857

48. Nakayama M, Beilke JN, Jasinski JM, Kobayashi M, Miao D, Li M,Coulombe MG, Liu E, Elliott JF, Gill RG, Eisenbarth GS. Priming andeffector dependence on insulin B:9-23 peptide in NOD islet autoimmunity.J Clin Invest 2007;117:1835–1843

49. Kobayashi M, Abiru N, Arakawa T, Fukushima K, Zhou H, Kawasaki E,Yamasaki H, Liu E, Miao D, Wong FS, Eisenbarth GS, Eguchi K. AlteredB:9-23 insulin, when administered intranasally with cholera toxin adjuvant,suppresses the expression of insulin autoantibodies and prevents diabetes.J Immunol 2007;179:2082–2088

50. Krishnamurthy B, Dudek NL, McKenzie MD, Purcell AW, Brooks AG,Gellert S, Colman PG, Harrison LC, Lew AM, Thomas HE, Kay TW.Responses against islet antigens in NOD mice are prevented by toleranceto proinsulin but not IGRP. J Clin Invest 2006;116:3258–3265



original article activation of insulin-reactive cd8 t...

Documents